172 HAROLD H. PLOUGH 



striking than the actual counts registered in the first half of the 

 next brood, but that it probably disappears entirely in the latter 

 half. It also shows that the decrease between the first and 

 second broods in the control series is a gradual one. 



Analysis of the temperature effect by counts at short intervals 



In view of this fact, it seemed reasonable to suppose that 

 approximately the actual number of eggs which are affected by 

 exposure to high temperatures for a given period could be de- 

 termined if the output of a female could be counted for a series 

 of intervals shorter than the usual ten day broods. In order 

 to meet this difficulty, the following method was devised. Fx 

 females from both control and heat-treated series were mated 

 and placed in shell vials, 1 inch by 4 inches, with a small amount 

 of food in the bottom. Each vial was marked with the serial 

 number of the pair and the date when the pair was placed in it. 

 These pairs were transferred to new vials after the females had 

 been allowed to lay for two days, and such a series was con- 

 tinued as long as eggs were laid. The total for each two day 

 interval of each series which had received the same treatment 

 was taken, the percentage of crossingover calculated, and the 

 average number of eggs hatched for the interval recorded. A 

 table constructed on the basis of these results gives a complete 

 picture of the variation in the percentage of crossingover through- 

 out the entire output of any female. In the case of females which 

 have been treated with high temperatures for a given period, 

 it shows just how soon the effect of that temperature is reflected 

 in an increase in crossingover, and just how long that effect 

 persists. Between the control and treated series the crossing- 

 over percentages at any given point are comparable almost 

 within the limits of the probable error, for they represent the 

 mean of two series from the same Pi parent, run at the same 

 time, and given the same food. In table 13 the results are given 

 for three separate series, all from the same Pi parents, the second 

 and third members of which received the heat treatment at differ- 

 ent times. Care was not taken in this group to change the pairs 



